Abstract
Present knowledge on metabolic perturbation prior to liver cancer is limited, with most studies unable to quantify the absolute concentrations of metabolites and lacking a focus on female-specific findings. We conducted a 1:1 matched nested case-control study (187 pairs of cases and controls) within the Shanghai Women's Health Study. A targeted metabolomics method was used to quantify 186 metabolites in plasma samples collected at recruitment. A multivariable conditional logistic regression model was utilized to estimate the odds ratio (OR) and 95% confidence interval (CI). Restricted cubic spline function was used to characterize the dose-response associations. A pathway analysis was conducted to identify the most relevant pathways. A metabolic score was calculated via a linear combination of metabolites with nonzero coefficients in the LASSO logistic regression model. After adjustment for potential confounders and correction for multiple testing, 27 metabolites were associated with liver cancer risk and a non-linear association was observed for glutamic acid. Primary bile acid biosynthesis and amino acid biosynthesis and metabolism were important pathways involved in the etiology. A metabolic score derived from 10 metabolites showed a positive linear association with liver cancer risk (OR(per standard deviation): 8.44, 95% CI: 4.09-17.39). The metabolic score significantly improved the predictive performance of the model based on established risk factors alone, which included age, excess body weight, smoking, alcohol use, and medical history of hepatitis. Our findings reveal female-specific metabolic perturbations prior to liver cancer diagnosis and contribute to a better understanding of the etiology and prevention of liver cancer in women.